Scaling Agent Learning via Experience Synthesis

• URL: http://arxiv.org/abs/2511.03773v2
• Date: Mon, 10 Nov 2025 05:02:36 GMT
• Title: Scaling Agent Learning via Experience Synthesis
• Authors: Zhaorun Chen, Zhuokai Zhao, Kai Zhang, Bo Liu, Qi Qi, Yifan Wu, Tarun Kalluri, Sara Cao, Yuanhao Xiong, Haibo Tong, Huaxiu Yao, Hengduo Li, Jiacheng Zhu, Xian Li, Dawn Song, Bo Li, Jason Weston, Dat Huynh,
• Abstract summary: Reinforcement learning can empower autonomous agents by enabling self-improvement through interaction.<n>But its practical adoption remains challenging due to costly rollouts, limited task diversity, unreliable reward signals, and infrastructure complexity.<n>We introduce DreamGym, the first unified framework designed to synthesize diverse experiences with scalability in mind.
• Score: 100.42712232390532
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: While reinforcement learning (RL) can empower autonomous agents by enabling self-improvement through interaction, its practical adoption remains challenging due to costly rollouts, limited task diversity, unreliable reward signals, and infrastructure complexity, all of which obstruct the collection of scalable experience data. To address these challenges, we introduce DreamGym, the first unified framework designed to synthesize diverse experiences with scalability in mind to enable effective online RL training for autonomous agents. Rather than relying on expensive real-environment rollouts, DreamGym distills environment dynamics into a reasoning-based experience model that derives consistent state transitions and feedback signals through step-by-step reasoning, enabling scalable agent rollout collection for RL. To improve the stability and quality of transitions, DreamGym leverages an experience replay buffer initialized with offline real-world data and continuously enriched with fresh interactions to actively support agent training. To improve knowledge acquisition, DreamGym adaptively generates new tasks that challenge the current agent policy, enabling more effective online curriculum learning. Experiments across diverse environments and agent backbones demonstrate that DreamGym substantially improves RL training, both in fully synthetic settings and in sim-to-real transfer scenarios. On non-RL-ready tasks like WebArena, DreamGym outperforms all baselines by over 30%. And in RL-ready but costly settings, it matches GRPO and PPO performance using only synthetic interactions. When transferring a policy trained purely on synthetic experiences to real-environment RL, DreamGym yields significant additional performance gains while requiring far fewer real-world interactions, providing a scalable warm-start strategy for general-purpose RL.

Related papers

• From Self-Evolving Synthetic Data to Verifiable-Reward RL: Post-Training Multi-turn Interactive Tool-Using Agents [23.583947864141162]
  EigenData is a hierarchical multi-agent engine that synthesizes tool-grounded dialogues together with executable per-instance checkers.<n>Building on the synthetic data, we develop an RL recipe that first fine-tunes the user model and then applies GRPO-style training.<n>Our results suggest a scalable pathway for bootstrapping complex tool-using behaviors without expensive human annotation.
  arXiv  Detail & Related papers  (2026-01-30T06:01:23Z)
• Sim-to-Real Transfer via a Style-Identified Cycle Consistent Generative Adversarial Network: Zero-Shot Deployment on Robotic Manipulators through Visual Domain Adaptation [1.0499611180329804]
  Deep Reinforcement Learning (DRL) compromises its industrial adoption due to the high cost and time demands of real-world training.<n>Virtual environments offer a cost-effective alternative for training DRL agents, but the transfer of learned policies to real setups is hindered by the sim-to-real gap.<n>This work proposes a novel domain adaptation approach relying on a Style-Identified Cycle Consistent Generative Adversarial Network (StyleID-CycleGAN or SICGAN), an original Cycle Consistent Generative Adversarial Network (CycleGAN) based model.
  arXiv  Detail & Related papers  (2026-01-23T11:48:15Z)
• Dyna-Mind: Learning to Simulate from Experience for Better AI Agents [62.21219817256246]
  We argue that current AI agents need ''vicarious trial and error'' - the capacity to mentally simulate alternative futures before acting.<n>We introduce Dyna-Mind, a two-stage training framework that explicitly teaches (V)LM agents to integrate such simulation into their reasoning.
  arXiv  Detail & Related papers  (2025-10-10T17:30:18Z)
• UserRL: Training Interactive User-Centric Agent via Reinforcement Learning [104.63494870852894]
  Reinforcement learning (RL) has shown promise in training agentic models that engage in dynamic, multi-turn interactions.<n>We propose UserRL, a unified framework for training and evaluating user-centric abilities through standardized gym environments.
  arXiv  Detail & Related papers  (2025-09-24T03:33:20Z)
• AgentGym-RL: Training LLM Agents for Long-Horizon Decision Making through Multi-Turn Reinforcement Learning [129.44038804430542]
  We introduce AgentGym-RL, a new framework to train LLM agents for multi-turn interactive decision-making through RL.<n>We propose ScalingInter-RL, a training approach designed for exploration-exploitation balance and stable RL optimization.<n>Our agents match or surpass commercial models on 27 tasks across diverse environments.
  arXiv  Detail & Related papers  (2025-09-10T16:46:11Z)
• WebEvolver: Enhancing Web Agent Self-Improvement with Coevolving World Model [55.276852838877346]
  Self-evolving agents are trained on trajectories sampled autonomously based on their own policies.<n>We propose a novel framework that introduces a co-evolving World Model LLM.<n>This world model predicts the next observation based on the current observation and action within the web environment.
  arXiv  Detail & Related papers  (2025-04-23T02:54:31Z)
• Learn-by-interact: A Data-Centric Framework for Self-Adaptive Agents in Realistic Environments [33.83610929282721]
  Learn-by-interact is a data-centric framework to adapt large language models (LLMs) to any given environments without human annotations.<n>We assess the quality of our synthetic data by using them in both training-based scenarios and training-free in-context learning (ICL)<n>Experiments on SWE-bench, WebArena, OSWorld and Spider2-V spanning across realistic coding, web, and desktop environments show the effectiveness of Learn-by-interact.
  arXiv  Detail & Related papers  (2025-01-18T22:34:41Z)
• A Platform-Agnostic Deep Reinforcement Learning Framework for Effective Sim2Real Transfer towards Autonomous Driving [0.0]
  Deep Reinforcement Learning (DRL) has shown remarkable success in solving complex tasks. transferring DRL agents to the real world is still challenging due to the significant discrepancies between simulation and reality. We propose a robust DRL framework that leverages platform-dependent perception modules to extract task-relevant information.
  arXiv  Detail & Related papers  (2023-04-14T07:55:07Z)
• Mastering the Unsupervised Reinforcement Learning Benchmark from Pixels [112.63440666617494]
  Reinforcement learning algorithms can succeed but require large amounts of interactions between the agent and the environment. We propose a new method to solve it, using unsupervised model-based RL, for pre-training the agent. We show robust performance on the Real-Word RL benchmark, hinting at resiliency to environment perturbations during adaptation.
  arXiv  Detail & Related papers  (2022-09-24T14:22:29Z)
• Cloud-Edge Training Architecture for Sim-to-Real Deep Reinforcement Learning [0.8399688944263843]
  Deep reinforcement learning (DRL) is a promising approach to solve complex control tasks by learning policies through interactions with the environment. Sim-to-real approaches leverage simulations to pretrain DRL policies and then deploy them in the real world. This work proposes a distributed cloud-edge architecture to train DRL agents in the real world in real-time.
  arXiv  Detail & Related papers  (2022-03-04T10:27:01Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.